278 Earth and Planetary Science Letters, 75 (1985) 278-288 Elsevier Science Publishers B.V., Amsterdam Printed in The Netherlands

[2]

Palaeomagnetic argument for a stationary Spitsbergen relative to the British Isles (Western Europe) since late Devonian and its bearing on North

Atlantic reconstruction

T.H. Torsvik ~, R. Lovlie 1 and B.A. Sturt 2

i Institute of Geophysics, University of Bergen, Bergen (Norway) 2 Institute of Geology, University of Bergen, Bergen (Norway)

Received March 14, 1985; revised version received June 18, 1985

Palaeomagnetic results from the Devonian of Central Spitsbergen (Mimer Valley Formation, Billefjorden) suggest an average palaeofield direction for the late Devonian/ear ly Carboniferous with declination = 227.5 °, inclination = - 30.6 °, corresponding to a relative pole position at lat. = 24°S, long. = 325°E. This magnetization which is post-tectonic is regarded as acquired chemically, linked with the Upper Devonian Svalbardian phase of deformation. This result, together with some recent data from Central Spitsbergen, defines a Devonian polar wander segment that accords extremely well with corresponding British data. We interpret this in terms of no major displacement of Central Spitsbergen as compared with Western Europe since Devonian time; Svalbard and the Barents Sea region have probably formed a fairly stable platform since the early Devonian. On the balance of available geological and palaeomagnetic evidence from Spitsbergen, the British Isles, and Newfoundland, the North Atlantic region does not appear to have been subjected to strike-slip motions in the order of thousands of kilometers in or since the late Devonian.

I. Introduction

The Iapetus Ocean was presumably initiated in late Precambrian time and bounded by passive margins until the early Palaeozoic. Subsequent plate convergence and marginal deformation, for- ming the Caledonian-Appalachian orogen, is first recognized in the Lower /Middle Ordovician, par- ticularly in the northern Appalachians, British Isles and Scandinavia [1,2]. Accompanying later exten- sive deformations and continental suturing in the mid-Palaeozoic, late Devonian /Carboni fe rous mega-shearing and continental re-arrangement, in- cluding the shear zones of Spitsbergen, northern Britain and Newfoundland/New England (Fig. 1A), have been proposed [3 8]. Due to suggested Devonian /ear ly Carboniferous palaeomagnetic discordances between the northeastern seaboard of North America (Acadia) and the interior of the continent, Kent and Updyke [9,10] argued that Acadia was originally situated some 1500 km south of the North American craton, and moved into its present relative position during the Carboniferous.

Van der Voo and Scotese [6] subsequently pos- tulated a ca. 2000 km early Carboniferous sinistral displacement along the Great Glen Fault (GGF), Scotland, in attempting to explain apparent palaeomagnetic latitudinal discrepancies between Europe and North America in the classical Bullard et al. [11] assemblage. In the latter model Acadia and Britain south of G G F were attached to the European margin of the shear zone, while Scotland north of the G G F was joined to the North American side.

Large-scale crustal displacement hypotheses have a tendency to be favoured and in turn adopted in the literature unless strong evidence is available to the contrary. In the present paper palaeogeo- graphical implications of some new palaeomagnetic data from the Devonian of Spitsbergen are dis- cussed with particular reference to the British Isles. Mega-shear hypotheses as put forward by Harland and co-workers [4,8,12-14] and Van der Voo and Scotese [7] are critically evaluated and discounted.

The almost undeformed Upper Silurian/ Devonian red sandstone and siltstone of the Wood

0012-821X/85/$03.30 © 1985 Elsevier Science Publishers B.V.

Bay Formation of Dicksonfjorden, Spitsbergen, have been studied palaeomagnetically by Storet- vedt [15] and Lovlie et al. [16]. Based on an overall scattered distribution of stable remanence direc- tions, almost coinciding azimuthal orientation of remanent directions and principal axes of maxi- mum susceptibility, and magnetomineralogical/ petrological observations, Lovlie et al. [16] de- scribed the magnetization as a detrital one (DRM) carried by haematite. On the other hand, Storet- vedt [15], working on beds apparently having a much larger proportion of cation-deficient phases (maghemites), suggested that the complex distribu- tion of remanence directions was due to multicom- ponent magnetization, brought about by post-de- positional chemical changes in a palaeofield of varying polarity.

Along with a recent palaeomagnetic field collec- tion in the Dicksonfjorden area (1982, 1983), sam- pling was also performed at five locations of the Middle Devonian Mimer Valley Formation (later referred to as the Billefjorden sediments). These sediments, which form the basis of the present discussion, are exposed in a small area im- mediately to the west of the Billefjorden linea- ment, and include grey-green sandstones, quart- zitic siltstones and mudstones [17].

2. Mid-Palaeozoic tectonic setting of Spitsbergen

The pre-Carboniferous geology of Spitsbergen is basically a distinction between the Upper Si lurian/Devonian Old Red Sandstone (ORS) and the Precambrian/Lower Palaeozoic Hecla Hoek Formation. The latter complex constitutes the area of main Caledonian metamorphism and of the late tectonic granite emplacement on Svalbard, for which an age ranging from 390 to 440 Ma [18-21] is suggested based on radiometric data. The Old Red Sandstone was deposited unconformably on the metamorphics of Central Spitsbergen and sub- sequently affected by late Devonian (Svalbardian) folding and deformation (notable along major fault zones). Spitsbergen includes numerous N - S / N W - S E structural lineaments that presumably played a controlling role in post-Caledonian sedi- mentation [22].

Palaeomobilistic models on the evolution of Svalbard include at least three major late De- vonian transcurrent fault zones (Fig. 1B). The

279

• .. vC~,oE~- ~,,~,AN ,ONE q E/~ ~ n A.EAS I~ r . z

~ ~/~ / I NORTH ATLANTIC

~ " f . ~ \'-",/ EA~LY AND ,S,ODLE 1 ? ALAEO O, / O.O,EN

z o N E

DtC K S O N I i l I ~ ~ E N l LATE SILURIAN AND DEVONIAN 5EDI MEN TS

Fig. 1. A. Sketch map of the North Atlantic bordering regions affected by the early and middle Palaeozoic orogenies, includ- ing suggested major fracture zones. Simplified after Harland [8] and Harland and Wright [13]. Diagram B shows the distribu- tion of the Old Red Sandstone graben of Central Spitsbergen bounded by the Billefjorden Fault Zone (BFZ) to the east. GSFZ: Greenland-Svalbard Fault Zone (postUlated); CWFZ: Central-West Fault Zone (postulated).

Central-West and Billefjorden Fault Zones (CWFZ and BFZ) are considered to define provincial boundaries within Spitsbergen, and based on com- parative facies analysis between Spitsbergen and other margins circumscribing the northern North Atlantic (Greenland, Arctic Ganada), they have been thought to involve considerable strike-slip movement [4,8,12-14]. Harland et al. [14] contend that left-lateral movements along the Billefjorden lineament occurred between East Spitsbergen and Greenland from Ordovician to Permian, and that the major part of the integral motion took place in two stages: (1) a late Devonian (Svalbardian) event of ca. 500 km, and (2) a mid-Carboniferous (late Mississippian) phase of ca. 1000 km or more. This would imply that the eastern province o f Spits-

280

bergen was originally situated south of Central Spitsbergen and later united by sinistral faulting. On the other hand, and in sharp contrast to the model of Harland and co-workers, recent struct- ural mapping of rock formations adjacent to the Billefjorden Fault Zone does not tend to support large-scale lateral movements (D. Douglas, per- sonal communicat ion, 1984).

3. Palaeomagnetic results from the Spitsbergen Old Red Sandstone

Measurements of anisotropy of magnetic sus- ceptibility (AMS) in the Billefjorden sediments

N

W E

PI 1.1"

5 B.

1.o5 • • "• /

• • ~ o o . •

1.05 1.1 Fig. 2. A. Stereographic presentation of principal axes of maximum (squares), intermediate (triangles) and minimum (circles) susceptibility for the investigated Billefjorden sedi- ments. Axial ratios are shown in a Flinn diagram (B) in which P1 is lineation and P3 foliation. Stereoplot conventions throughout this paper are closed (open) symbols for downward (upward) pointing inclinations.

reveal bedding-parallel, steeply inclined magnetic foliation planes striking NNE-SSW, giving both prolate and oblate magnetic ellipsoids (Fig. 2). Total anisotropy (Kmax/Kmin) is typically around 5%. The fairly good grouping of Kma x (lineation : fold axis oriented) may be an original feature, but is more likely to have been imposed during defor- mat ion and folding (pencil cleavage). The mag- netic fabric data is consistent with structural field evidence suggesting W N W - E S E compression.

Curie temperatures (T~) around 680°C and the mode of isothermal remanent ( IRM) acquisition versus increasing magnetizing field, suggest that haematite dominates the bulk magnetic properties (Fig. 3). A secondary phase with T. around 560-580°C (probably magnetite) is readily pro- duced during heating above 600°C. Magnetite pro- duction on thermal treatment is also suggested by a 2 -3 times increase in bulk susceptibility, and by a general increase in low-coercivity material and total IRM intensity.

IRM B. 40A/m

k\ Dr20 °C)

\ \ \

\ \ A . \ \

\

~ ,, "', o:1 o.2 o.3 ( T }

, T°CxlO0

Fig. 3. Example of "saturation" magnetization versus tempera- ture (diagram A) showing formation of a new magnetic phase ("magnetite") after heat treatment to ca. 700°C (cf. cooling curve). Diagram B gives an example of IRM acquisition versus increasing field. Note the huge difference between the unheated and heated rock sample, demonstrating the production of a secondary spinel phase.

320,Up. 2mA/m

NRiq zoo" C

~(~570 B 7 "~"0 600"C

230,230

A major part of the rock collection had natural remanent magnetizations (NRM) lower than the instrumental noise level (ca. 0.5 mA/m), hence, only 18 samples were subjected to thermal demag- netization (NRM intensities typically in the 2-4 mA/m range). Stepwise thermal demagnetization has only uncovered one component of magnetiza- tion as evidenced from optimal vector projections (Fig. 4). Though a fair proportion of the magnetic moment frequently remains after demagnetization to 600°C, it proved impossible to obtain sensible results at higher temperatures. This erratic stage is attributed to the formation of secondary magnetite as discussed above. In the NRM-600°C range, however, all tested samples designate linear trajec- tories towards the origin of the vector plots. From a total of 18 tested samples (those having NRM

6 ~ soo'c NRM \

" ,~5 5 0

"0 600"C

220~220

NRH 3 3 0~Up / 8. 3oo°c ,,, "0. -C~,~O0

bJoo

"c& 5 7 o -. 60o'c B 2 9

"~"O

¢ -" w 2409240

3~o, uP l,.

B 3 2

Fig. 4. Orthogonal vector diagrams showing examples of rema- nence behaviour on thermal demagnetization. The plots are optimal versions, and open (closed) symbols denote projections on the vertical (horizontal) plane. Note that the characteristic components of magnetization are not significantly different from the direction of natural remanent magnetization (NRM).

Fig. 5. Distribution of stable remanence directions for all tested samples of the Billefjorden sediments, before (A) and after (B) structural unfolding.

intensities above the instrument noise level) 14 samples form a very well-defined grouping of re- manence directions, having southwest declinations and upward (negative) inclinations of around 30 ° prior to structural unfolding (Fig. 5A). Bedding correction (Fig. 5B) yields almost similar declina- tions but downward-dipping inclinations. Unfor- tunately, the area concerned has a fairly uniform orientation of bedding so that a fold test, to dif- ferentiate between post(/syn)- and pre-folding origin of magnetization, cannot be accomplished.

The Dicksonfjorden Old Red Sandstone succes- sion (Wood Bay Formation), located towards the central part of the ORS graben of Spitsbergen, in general shows scattered directions of stable rema- nent magnetization [15,16]. For the investigated beds Lovlie et al. [16] ascribe the scatter primarily to randomizing effects from high-energy environ- ments acting during accumulation (including detri- tal haematite). However, three rock sections show reasonably grouped directions of magnetization

281

282

TABLE 1 Overall palaeomagnetic results from the Billefjorden Old Red Sandstone (pre~ent study) along with re-calculated/selected results from the Dicksonfjorden ORS [16]

Formation D (o) I (o) N a95 (o) K Pole dp/dm Mimer Valley Fm. Bille- 227.5 - 30.6 18 8.7 16.8 24°S,325°E 5 /10

fjorden; without tect. corr. Mimer Valley Fm. Bille- 227.4 + 29.8 18 8.7 16.8 8°N,331°E 5 /10

fjorden; tect. corrected Wood Bay Fm., Dicksonfjorden

Group 1 231 + 22 21 12 6.5 4°N,325°E 7 /13 Group 2 049 + 15 13 16.3 5.7 15°S,325°E 9 /17 Group 3 239 +6 9 10.6 19.8 3°S,317°E 5/11

D = mean declination, I = mean inclination, N = number of specimens, and a95 and K are the semi-angle of the 95% confidence cone around the mean direction and the precision parameter, respectively [54]. dp and dm are the semi-axis of the oval of 95% confidence around mean pole position, along the palaeomeridian and perpendicular to it, respectively.

(cf. [16, fig. 11]). The corresponding mean direc- tions (Table 1), excluding some aberrant specimen directions, are plotted in Fig. 6 along with the average Billefjorden results, i.e. before (B-UC) and after (B-TC) correction. It is noted that all these results define the same palaeomeridian, but the Dicksonfjorden magnetizations have inclinations between the two Billefjorden directions (B-TC and B-UC).

Fig. 6. Stereoplot showing the Billefjorden mean direction, before (B-UC) and after (B-TC) tectonic correction, in conjunc- tion with directions for the most well-grouped populations of the Dicksonfjorden ORS sediments, groups DI-I I I .

4. Comparison with British results

In Europe the British palaeomagnetic data de- fine the most coherent apparent polar wander (APW) path for the Palaeozoic, though details in the polar pattern are interpretated differently by various authors. From palaeomagnetic data, minor left-lateral displacement has been suggested along the Great Glen Fault [29], but at least to a first approximation it is fair to say that the polar paths north and south of the Great Glen Fault (GGF) show no significant differences [23]. Also, recent palaeomagnetic data from the ca. 400-430 Ma "newer granites" [24-26], situated on both sides of the GGF, are shown to be consistent, thus dis- puting the large-scale post-Devonian movement advocated by Van der Voo and Scotese [6]. From at least post-Silurian palaeornagnetic data Scot- land north of GGF is clearly "European", and the fault cannot seriously be considered as a mega- shear of say > 500 km in the mid-Palaeozoic, in agreement with geological evidences [28].

The Lower Devonian lavas, Middle/Upper De- vonian sediments and volcanics (all three Scottish rocks), and British Permo-Carboniferous lavas and dykes reveal internally consistent data, suggesting rapid APW in this periode. Data from continental Europe are in general agreement with the British results, but the data, notably the Devonian ones, are more scattered and are clearly in need of laboratory re-examination. The equatorial VGP position held by the Lower Devonian lavas from Scotland is generally absent in investigated Lower

ORS sediments from Europe. This variation is most likely due to differences in the ability to retain the original Lower Devonian field compo- nent: the lavas probably acquired a stable Lower Devonian magnetization through an early stage of extensive oxidation (partly of deuteric origin) while the sediments in most cases have undergone a long-lasting magnetization history, including di-

A .

LOWER DEVONIAN

\ M-U D E V O N I A N , ~ C S ~ _ ~ C

BRITISH SCES \ I

B .

15 ° E

t I5ON EQ, UA 1"OR

- - -15o5

~ i - 30°5

300°E 315OE 330°E 345°E 0 o 15°E

D1 D S - TC

I I/ ~[ I D 311

SPITSBERGEN ~

~ I 5 ° N EQ,UA TOR

_ _ _ _ 1 5 ° S

30os

Fig. 7. The relative polar path for the British Isles. A. Devonian to Permian time. Lower Devonian: open triangles; Midd le / Upper Devonian: open circles; Upper Carboniferous/Permian: open squares. See Table 2 for formation codes. B. Spitsbergen poles in the framework of the British mid-late Palaeozoic APW path. Poles B-UC and B-TC are based on the Billefjorden ORS sediments, before and after tilt correction, respectively. Poles D 1 - 3 represent the Devonian strata of Dicksonfjorden. Fur- ther details in Table 1 and in the text.

283

agenetic processes, continuing in many cases per- haps throughout the Devonian. Furthermore, late Caledonian and younger magnetic overprinting are of importance in central Europe, and e.g. within the Armorican Massif, no Devonian palaeomag- netic data are available [30].

Fig. 7B despict the Spitsbergen Devonian data in the framework of the British Middle/Upper Palaeozoic (Devonian/Permian) results (Table 2). From this comparison, the Billefjorden magnetiza- tion has two possible age alternatives: (1) the magnetization is pre-folding (initial) and of Lower Devonian age (cf. pole B-TC), or (2) the magneti- zation is post-tectonic (pole B-UC), i.e. remag- netized at around Upper Devonian time (Sval- bardian).

The fairly extensive tectonic deformation (in- cluding pronounced cleavage) of the investigated Billefjorden rocks, and the fact that these rocks are most likely Middle Devonian in age clearly favour the remagnetization alternative. Although the Billefjorden data represent a relatively small num- ber of tested samples, they all exhibit high-quality single-component magnetizations compared with

TABLE 2

Relevant Palaeozoic palaeomagnetic poles from the British Isles used in an APW master curve for comparison with other regions

Pole Ref. Notation in Fig. 7A

Lower Devonian Lorne Plateau Lavas 2°N, 321°E [31] LP Garabal Hill-Glen Fyne 5°S, 326°E [32] GH Arrochar Complex 8°S, 324°E [32] AC Midland Valley lavas/sed. 4°S, 320°E [33] MV Midland Valley lavas 2°N, 318°E [34] SL Cheviot Hill lavas/grani tes I°N, 329°E [35] CH

Middle/Upper Devonian Orkney lavas 24°S, 330°E [36] OL Hoy lavas " 23°S, 326°E [37] RH Caithness sst. 27°S, 329°E [38] CS Duncansby Neck 24°S, 329°E [39] DN John O'Groats sst. 24°S, 325°E [40] JG Foyers ORS sed. 30°S, 327°E [41] FS Shetlandignimbrites 20°S, 315°E [27] SI

Upper Carboniferous~Permian Exeter lavas 46°S, 345°E [42] EL Wackerfield dyke 49°S, 349°E [43] WD Whin Sill 44°S, 339°E [44] WS Peterhead dyke 41°S, 342°E [26] PD

284

the Dicksonfjorden sediments [16]. The Sval- bardian thermal and chemical environments were probably at their maximum in rocks adjacent to the Billefjorden lineament. Thus, the rocks of the present study (sampled close to the fault) appear to have a single-component magnetization due to complete magnetic resetting in Svalbardian time. The Dicksonfjorden results referred to here (D1-D3) are most likely of depositional origin and must therefore be older than the Billefjorden remanence. Poles D1 and D3 are seen to match the Lower Devonian pole for Britain extremely well. The suggested inclination error in the Dick- sonfjorden beds [16] may in fact be negligible due to the apparent equatorial latitudes of Central Spitsbergen in the Lower Devonian. The low palaeolatitude of Spitsbergen in the Devonian is also supported by palaeoclimatic [22] evidence (Fig. 8).

Thus, with reference to Fig. 7B, available Palaeozoic palaeomagnetic data from Central Spitsbergen may tentatively be interpreted as fol- lows: (1) Dicksonfjorden sediments (poles D1-

D3) record Lower Devonian compatible poles (de- positional age; magnetization probably unbiased by later orogenic episodes), and (2) Billefjorden sediments (pole B-UC) are considered to have late Devonian or possible early Carboniferous mag- netic ages. However, we emphas ize that the polar pattern pattern as outlined here is preliminary and actual magnetic ages are uncertain. Of the poles concerned, the inferred late Devonian ones seem to be most reliable. Nevertheless, the presently available results from Central Spitsbergen are re- markably consistent with corresponding British data, suggesting a fairly coherent tectonic rela- tionship in late/post-Devonian times.

5. Palaeozoic latitudinal discordances between Europe (Acadia) and North America?

The proposed Carboniferous (Mississippian) phase of strike-slip (ca. 1000 km) motion within Spitsbergen [14] is exclusively based on mega-shear models from elsewhere in the North Atlantic re- gion, where palaeomagnetic data have been used

A . B

CO 80

~ 60- k

t.u 20-

~ o .

,do 2bo 3bo M~ R. ~o

,2

V

"IF "III U

. . . . I

LATE DE~ ONIAN - EA&

q

j

LY CARBONIF ~ROUS

D 17*_5ON 15ON

E ~ R

Y GREY MARINE TO DELTAIC CLASTICS WITH COALS I F NONE-EVAP. CARBS & CLASTICS

I I I CARBONATES~EVAPORIFES & RED CLASTICS I I GREY FLUVIAL CLASTICS & COAL I OLD RED SANDSTONE

Fig. 8. The palaeolatitude/palaeoclimatic migration of Svalbard (A), redrawn from Steel and Worsley [22], in conjunction with the palaeomagnetically based latitudinal configuration (for Central Spitsbergen, Scandinavia and the British Isles) in Middle Devonian/ early Carboniferous (B). The palaeolatitude of Central Spitsbergen is based on the inferred Middle Devonian age Billefjorden magnetization. Note the good correspondence between the palaeoclimatically and palaeomagnetically based latitudes.

for a rguments that the coasta l C a n a d i a n Mar i - t i m e - N e w England region [9,10] and la ter in- c luding the Ava lon Plat form, East N e w f o u n d l a n d [47], was or ig inal ly s i tuated some 15 -20 ° far ther south, the present j ux t apos i t i on of the coasta l re- gion (Acadia) and cra tonic N o r t h Amer i ca be ing b r o u g h t abou t by mega-shear ing in mid- la te Carboni fe rous . Van der Voo and Scotese [6] have ex tended this idea, p ropos ing lef t - la teral displace- men t in tlae o rder of 2000 km between cra tonic N o r t h Amer i ca and Europe. Van der Voo and Scotese [6] suggested that the Grea t Glen fault, Scot land, to be the app rop r i a t e cand ida t e for such a ma jo r crusta l t ransla t ion, but it has subsequent ly been po in ted out [24,25,27,38] that pa laeomagne t i - cal ly nor thern Scot land is c lear ly European in pos t -S i lur ian t ime, and the avai lable da t a are to- ta l ly i ncompa t ib l e with the p r o p o s e d mega-scale t ranscurrence.

C o m p a r e d with the British Isles, pa laeomagne t i c da t a from the N o r t h Amer ican cra ton [7,46] show a lmost no A P W in D e v o n i a n / P e r m i a n t ime (Fig. 9B). This m a y be taken as evidence for severe magnet ic reset t ing in the Hercyn ian [45,46,55-57]. Fur the rmore , I rving and Strong [48,49] have shown that results f rom the early Carbon i fe rous of west-

B. 315OE 345~E 0 o 15OE

{ ' " - , E(2UATOR

~" " -CL 30°5 - -.OG 2U

A.

~ -- I5ON U~TOR

15o5

- - 30o5

C Fig. 9. Middle/Upper Palaeozoic palaeomagnetic poles from Newfoundland (A) and cratonic North America (B). The poles have been rotated according to the Bullard et al. [11] continen- tal configuration for comparison with the British (European) apparent polar wander path (open arrows). Lower Carbonifer- ous rocks from both western Newfoundland, DG [49], and eastern Newfoundland, TF [47], are characterized by Upper Carboniferous/Permian magnetic overprinting (DG-CU and TF-CU, respectively). Poles DG-CL and TF-CL are thought to represent the original magnetization of these rocks. B. Mean poles [7] for the Devonian (DM-DU), Carboniferous (CL, CU) and Permian (PL, PU) respectively. Note the lack of systematic polar migration for cratonic North America as compared with that for the British Isles.

285

ern and eas tern N e w f o u n d l a n d are very close (Fig. 9A), imply ing that no large-scale str ike-sl ip mot ion has occurred between Acad i a and cra tonic N o r t h Amer i ca since ear ly Carboni ferous . The la t ter au thors have also demons t r a t ed the widespread late Palaeozoic magnet ic overpr in t ing in New- foundland , an observa t ion that accords comple te ly with the pa laeomagne t i c evidence f rom nor thern Scot land [24,37,39,40,50,51]. In fact, it was the

/ -I /_ E A/RLy } 1 \ \ ICARBONIFEROUS \ - - 15°A~ / L , T ~ I I t E CARBONIFEROU L --

f ,

EARLY PERMIAN \

_ _ EClUA T O R

5 ° S

Fig. 10. Late Devonian/early Carboniferous (A) and late Carboniferous/Permian (B) reconfiguration in the North Atlantic based on presently available palaeomagnetic data. Longitudinal positioning which for diagram B is according to the Bullard et al. [11] fit is only optional. Reference poles for latitudinal positioning are according to Figs. 8 and 9. Scaodinavia/coast-line Central Europe, and Spitsbergen in the late Carboniferous/Permian, is with reference to British data (assuming no relative movements). The northerly position of the North American craton in late Devonian/early Carbonifer- ous time is thought to be incorrect, reflecting substantial mag- netic resetting of mid-Palaeozoic rocks during the Hercynian orogeny. The arrow indicates the assumed true position of cratonic North America relative to Acadia/Newfoundland.

286

strong"Permian" overprint that Van der Voo and Scotese erroneously accepted as the original mag- netization in Orcadian basin rocks, which led to their artificial tectonic interpretation.

Two palaeo-reconstructions, revealing the con- troversial transition between late Devonian/ear ly Carboniferous and late Carboniferous/early Per- mian, are outlined in Fig. 10A and B. These reconstructions are strictly latitudinal. They are tentatively corrected for opening of the North Atlantic using the classical Bullard et al. [11] as- semblage (apart from cratonic North America in Fig. 10A for which longitude versus Acadia / Newfoundland is optional), and they summarize the present status in the North Atlantic transcur- rence debate. In Fig. 10A the North American craton (based on present palaeomagnetic data) is positioned north of Acadia and Europe. New- foundland forms one single unit, and the whole of Scotland is seen as part of Europe. However, there are reasons to believe that the relative latitude of the North American craton in late Devonian/ear ly Carboniferous is incorrect, caused by Permian magnetic resetting. Thus, Irving and Strong [48] argue that regions of the craton, apparently free from Kiaman overprinting, should reveal palaeo- latitudes that accord with those of Acadia and Europe. Consequently, the late Devonian/ear ly Carboniferous configuration of cratonic North America and Europe was apparently similar to the one for late Carboniferous/early Permian (Fig. 10B), and mega-shearing is not required. If so, a Mississippian phase of strike-slip motion within Spitsbergen [14] is not supported by palaeomag- netic data outside the Arctic Caledeonides.

6. Conclusion

Palaeomagnetic results from the Devonian Billefjorden sediments (Mimer Valley Formation) define a magnetic meridian compatible with recent results from the Dicksonfjorden Old Red Sand- stone. The investigated rocks have suffered later deformation and folding, and well-defined rema- nence directions similar to data from the mid-late Devonian of the British Isles suggest remagnetiza- tion linked with a late Devonian (Svalbardian) phase of deformation. The Dicksonfjorden ORS was probably not significantly affected by the Svalbardian orogenic disturbance. However, the

palaeomagnetic reliability of the Dicksonfjorden is uncertain (inclination error), though it is not un- likely that the magnetization is fairly accurate and records a stable depositional period in the Lower Devonian.

In Middle/late Devonian ( /ear ly Carbonifer- ous) the British Isles and southern Norway were apparently in equatorial areas (Fig. 8B), while Central Spitsbergen was confined to a ca. 15 ° northerly latitude, a conclusion which is sustained by palaeoclimatic evidence from Spitsbergen (Fig. 8A). Palaeomobilistic views given by Harland [8] and Harland and Wright [13], separating Spits- bergen by several major shear zones, cannot be fully tested by the available palaeomagnetic data, but it seems likely .that at least Central Spitsbergen has not been involved in major displacement as compared with Europe, which in turn may reflect the fact that Svalbard and the Barents Sea region have formed a stable platform since late De- vonian/ear ly Carboniferous time. If one accepts the Dicksonfjorden results, it implies, in fact, that Central Spitsbergen has remained fixed relative to Europe since Lower Devonian.

The mega-shear argument has been employed to explain differences in the various tectonic zones of Newfoundland, though recent palaeomagnetic data conclusively show that eastern and western Newfoundland were attached to each other and to the North American craton (due to its litho- tectonic relationship with western Newfoundland) prior to early Carboniferous [49], and presumably since the Acadian orogeny in Middle Devonian. The proposal of Van der Voo and Scotese [6] of a mega-shear along the Great Glen fault can simi- larly be discounted. We submit, that on balance of both geological and palaeomagnetic evidence from Spitsbergen, the British Isles, and Newfoundland, that mega-scale transcurrence in the Nbrth Atlantic region during the mid-late Palaeozoic is not proven and indeed must be regarded as highly suspect. Strike-slip controlled Old Red Sandstone basins within the Caledonian Fold Belt may certainly have occurred locally [3,22,53], but do not imply or prove mega-shear in the order of thousands of kilometers.

Acknowledgements This research project has been supported by the

Norwegian Research Council for Science and the

287

Humanities. Considerable field assistance pro- vided by the Norwegian Polar Institute, and com- ments from E.R. Deutsch, R. Hyde and D.F. Strong are gratefully acknowledged. Criticism and improvement of the manuscript given by K.M. Storetvedt, R. Van der Voo and two other anony- • mous referees are greatly appreciated.

Norwegian Lithosphere Project Contribution No. (02).

References

1 B.A. Sturt, The accretion of ophiolitic terrains in the Scandinavian terrains, in: Ophiolites and Ultramafic Rocks, H.J. Zwart, P. Hartman and A.C. Tobi, eds., Geol. Mijnbouw 63, 201, 1984.

2 H. Williams, Miogeosynclines and suspect terranes of the Caledonian-Appalachian Orogen: tectonic patterns in the North Atlantic region, Can. J. Earth Sci. 21,887, 1984.

3 J. Hailer, Geology of the East Greenland Caledonides, Interscience, London, 1971.

4 W.B. Harland and R.A. Gayer, The Arctic Caledonides and earlier oceans, Geol. Mag. 109, 289, 1972.

5 W.A. Morris, Transcurrent motion determined palaeomag- netically in the Northern Appalachian and Caledonides and the Acadian orogeny, Can. J. Earth Sci. 13, 1236, 1976.

6 R. Van der Voo and C. Scotese, Palaeomagnetic evidence for a large (ca. 2000 km) sinistral offset along the Great Glen Fault during Carboniferous time, Geology 9, 583, 1981.

7 R. Van der Voo, Palaeomagnetic constraints on the as- semblage of the Old Red Continent, Tectonophysics 91, 271, 1983.

8 W.B. Harland, Contribution of Spitsbergen to understand- ing of tectonic evolution of North Atlantic region, Am. Assoc. Pet. Geol. Mem. 12, 817, 1969.

9 D.V. Kent and N.D. Opdyke, Palaeomagnetism of the Devonian Catskill Red Beds: evidence for motion of the coastal New England-Canadian Maritime region relative to Cratonic North America, J. Geophys. Res. 83, 4441, 1978.

10 D.V. Kent and N.D. Opdyke, The early Carboniferous palaeomagnetic field of North America and its bearing on tectonics of the Northern Appalachians, Earth Planet. Sci. Lett. 44, 365, 1979.

11 E.C. Bullard, J.E. Everett and A.G. Smith, The fit of the continents around the Atlantic, Philos. Trans. R. Soc. London, Ser. A 258, 41, 1965.

12 W.B. Harland, The tectonic evolution of the Arctic-North Atlantic region, Philos. Trans. R. Soc. London, Ser. B 258, 59, 1965.

13 W.B. Harland and N.J.R. Wright, Alternative hypothesis for the pre-Carboniferous evolution of Svalbard, Norsk Polarinst. Skr. 167, 90, 1979.

14 W.B. Harland, B.A. Gaskell, A.P. Haefford, E.K. Lind and P.J. Perkins, Outline or Arctic post-Silurian continental displacements, in: Petroleum Geology of the North European Margin, Graham and Trotman, eds., pp. 137-148, 1984.

15 K.M. Storetvedt, Old Red Sandstone palaeomagnetism of Central Spitsbergen and the Upper Devonian (Svalbardian) phase of deformation, Norsk Polarinst. Arb. 59, 1972.

16 R. Lovlie, T. Torsvik, M. Jelenska and M. Levandowski, Evidence for detrital remanent magnetization carried by hematite in Devonian Red Beds from Spitsbergen; palaeomagnetic implications, Geophys. J. R. Astron. Soc. 79, 573, 1984.

17 P.F. Friend, The Devonian stratigraphy of north and central Vest-Spitsbergen, Proc. Yorks. Geol. Soc. 28, 77, 1961.

18 R.A. Gayer, D.G. Gee, W.B. Harland, J.A. Miller, H.R. Spall, R.H. Wallis and T.S. Winsnes, Radiometric age de- terminations on rocks from Spitsbergen, Norsk Polarinst. Skr. 137¢ 39 pp., 1966.

19 A.A. Krasil'scikov, Stratigraphy and paleotectonics of the Precambrian and early Palaeozoic of Spitsbergen, Trans., Sci. Res. Inst. Geol. Arctic 172, 170 pp., 1973.

20 M.G. Ravich, Is there an early Precambrian granite-gneiss complex in Northwestern Spitsbergen?, Skr. Norsk Polarinst. 167, 9, 1979.

21 S.A. Abakumov, Peculiar features of regional metamor- phism of Northwestern Spitsbergen, Skr. Norsk Polarinst. 167, 29, 1979.

22 R.J. Steel and D. Worsley, Svalbard's post-Caledonian strata--an atlas of sedimentological patterns and palaeo- geographic evolution, in: Petroleum Geology of the North European Margin, Grahm and Trotman, eds., pp. 59-69, 1984.

23 J.C. Briden, H.B. Turnell and D.R. Watts, British palaeomagnetism, Iapetus Ocean and the Great Glen Fault, Geology 12, 428, 1984.

24 T.H. Torsvik, R. Lovlie and K.M. Storetvedt, Multicompo- nent magnetization in the Helmsdale granite, North Scot- land; geotectonic implication, Tectonophysics 98, 111, 1983.

25 T.H. Torsvik, Palaeomagnetism of the Foyers and Strontian granites, Scotland, Phys. Earth Planet. Inter. 36, 163, 1984.

26 T.H. Torsvik, Palaeomagnetic results from the Peterhead granite, Scotland; implication for regional late Caledonian magnetic overprinting, Phys. Earth Planet. Inter. (in press).

27 K.M. Storetvedt and T.H. Torsvik, Palaeomagnetic results from Esha Ness Ignimbrite, Shetland, Phys. Earth Planet. Inter. 37, 169, 1985.

28 D.I. Smith and J. Watson, Scale and timing of movements on the Great Glen Fault, Scotland, Geology 11,523, 1983.

29 K.M. Storetvedt, A possible large-scale sinistral displace- ment along the Great Glen Fault in Scotland, Geol. Mag. 111, 23, 1974.

30 H. Perroud, M. Robardet, R. Van der Voo, N. Bonhommet and F. Paris, Revision of the age of magnetization of the Montmartin red beds, Normandy, France, Geophys. J. R. Astron. Soc. 80, 541, 1984.

31 A.G. Latham and J.C. Briden, Palaeomagnetic field direc- tions in Siluro-Devonian lavas of the Lorne Plateau, Scot- land, and their regional significance, Geophys. J. R. Astron. Soc. 43, 243, 1975.

32 J.C. Briden, Palaeomagnetic results from the Arrochar and Garabal Hill-Glen Fyne igneous complexes, Scotland, Ge- ophys. J.R. Astron. Soc. 21,457, 1970.

33 J.T. Sallomy and J.D.A. Piper, Palaeomagnetic studies in the British Caledonides, IV. Lower Devonian lavas of the

288

Strathmore region, Scotland, Geophys. J. R. Astron. Soc. 34, 47, 1973.

34 T.H. Torsvik, Magnetic properties of the Lower Old Red Sandstone lavas in the Midland Valley; palaeomagnetic and tectonic consideration, Phys. Earth Planet. Inter. (in press).

35 L. Thorning, Palaeomagnetic results from Lower Devonian rocks of the Cheviot Hills, Northern England, Geophys. J. R. Astron. Soc. 36, 487, 1974.

36 K.M. Storetvedt and N. Petersen, Palaeomagnetic proper- ties of the Middle-Upper Devonian volcanics of the Orkney Islands, Earth Planet. Sci. Lett. 14, 269, 1972.

37 K.M. Storetvedt and A.H. Meland, Geological Interpreta- tion of palaeomagnetic results from Devonian rocks of Hoy, Orkney (submitted to Scott. J. Geol.).

38 K.M. Storetvedt and T.H. Torsvik, Palaeomagnetic re-ex- amination of the basal Caithness Old Red Sandstone; aspects of local and regional tectonics, Tectonophysics 98, 151, !983.

39 K.M. Storetvedt, C.M. Carmichael, A. Hayatsu and H.C. Palmer, Palaeomagnetism and K/Ar results from the Duncansby volcanic neck, northeastern Scotland: superim- posed magnetizations, age of igneous activity and tectonic implications, Phys. Earth Planet. Inter. 16, 379, 1978.

40 K.M. Storetvedt and C.M. Carmichael, Resolution of super- imposed magnetization in the Devonian John O'Groats Sandstone, N. Scotland, Geophys. J. R. Astron. Soc. 58, 769, 1979.

41 S.J. Kneen, The Palaeomagnetism of the Foyers Plutonic Complex, Invernesshire, Geophys. J. R. Astron. Soc. 32, 53, 1973.

42 J.D. Cornwell, Palaeomagnetism of the Exeter Lavas, De- vonshire, Geophys. J. R. Astron. Soc. 12, 181, 1967.

43 D.H. Tarling, J.G. Mitchell and H. Spall, A palaeomagnetic and isotopic age for the Wackerfield Dyke of Northern England, Earth Planet. Sci. Lett. 18, 427, 1973.

44 K.M. Storetvedt and A. Gidskehaug, The magnetization of the Great Whin Sill, Northern England, Phys. Earth Planet. Inter. 2, 105, 1969.

45 J.L. Roy and W.A. Morris, Palaeomagnetic results from the Carboniferous of North America; Development of a chro- nostratigraphic Marker horizon for Canadian East Coast

Carboniferous strata, C.R. 9th Int. Congr. Carb. Strat. Geol., p. 23, 1979.

46 J.L. Roy, E. Tanczyk and P. Lapointe, The palaeomagnetic record of the Appalachians, in: Regional Trends of the Appalachian-Caledonian-Hercynian-Maurit amide Orogen, P.E. Schenk, ed., pp. 11-26, D. Reidel, Dordrecht, 1983.

47 D.V. Kent, Paleomagnetic evidence for post-Devonian dis- placement of the Avalon Platform (Newfoundland), J. Geo- phys. Res. 87, 8709, 1982.

48 E. Irving and D.F. Strong, Evidence against large-scale Carboniferous strike-slip faulting in the Appalachian- Caledonian orogen, Nature 310, 762, 1984.

49 E. Irving and D.F. Strong, Palaeomagnetism of the early carboniferous Deer Lake Group, western Newfoundland: no evidence for mid-Carboniferous displacement of "Acadia", Earth Planet. Sci. Lett. 69, 379, 1984.

50 D.H. Tarling, R.N. Donovan, J. Abou-Deeb and S.I. Batrouk, Palaeomagnetic dating of haematite genesis in Orcadian basin sediments, Scott, J. Geol. 12, 125, 1976.

51 P. Turner, Remanent magnetism of middle ORS lacustrine and fluviatile sediments from the Orcadian basin, Scotland, J. Geol. Soc. London 133, 37, 1977.

52 J.L. Roy, Paleomagnetism of Siluro-Devonian rocks from eastern Maine: discussion, Can. J. Earth Sci. 19, 225, 1982.

53 R.J. Steel, Late orogenic Devonian basin formation in the western Norwegian Caledonides, Geol. Surv. Can. Pap. 78, 57, 1979.

54 R.A. Fisher, Dispersion on a sphere, Proc. R. Soc. London, Ser. A 217, 295, 1953.

55 K.V. Rao, M.K. Seguin and E.R. Deutsch, Palaeomagne- tism of Siluro-Devonian and Cambrian granitic rocks from the Avalon Zone in Cape Island, Nova Scotia, Can. J. Earth Sci. 18, 1187, 1981.

56 J.L. Roy and P. Anderson, An investigation of the rema- nence characteristics of three sedimentary units of the Silurian Marcarene Group of New Brunswick, Canada, J. Geophys. Res. 86, 6351, 1981.

57 J.L. Roy and W.A. Morris, A review of palaeomagnetic results from the Carboniferous of North America; the con- cept of Carboniferous geomagnetic field horizon markers, Earth Planet. Sci. Lett. 65, 167, 1983.